reveal the signal
your private intelligence gets named before it gets automated
pillarssovereignty · navigation (inward)
artifactoperator graph
curriculum · five pillars · public
Five disciplines. Or you get quietly replaced.
The age of AI is not a tool problem. It is an operator problem. The five pillars are the disciplines that compound under that pressure instead of being deleted by it. The pillars are public. The rubrics are public. The essays are public. The install is the room where they get installed against your live business in twenty-one days.
public curriculum
- 01 sovereignty
- 02 systems
- 03 coordination
- 04 navigation
- 05 translation
twenty-one day install
- week 01 operator graph
- week 02 ai operating stack
- week 03 launched edge
five pillars · five essays · five rubrics · free
why five
AI does not replace operators. It replaces operators whose private intelligence was never named.
An LLM is a public graph compressed into weights. It knows what everyone knows. It does not know you — your taste, your refusals, your standards, your decision tree, the way you read a room. The operators getting replaced are not the bad ones. They are the smart ones whose private intelligence was never written down.
The curriculum names that private intelligence in five dimensions. Five pillars. One operator. Each pillar is a discipline. Each discipline has a public rubric. Each rubric is something you can mark your own work against today, alone, for free.
Sovereignty. Systems literacy. Coordination. Navigation. Translation. These are not the only five. They are the five we found load-bearing — the ones that, when missing, are the part that gets replaced.
The install is the room where the pillars get installed. The curriculum is the standard they get installed to.
the five disciplines
What the curriculum is.
Click any pillar to open it. Each discipline carries a failure mode, what gets installed, the public rubric, and the essay that goes deeper.
01 sovereignty your voice stays yours under load.
the failure
The model writes the email. The email goes out. You realise three weeks later you have been quietly saying things you do not actually believe — and your team has started believing them. The recovery is not writing better prompts. It is relearning what you mean before the model learns to sound like you.
what gets installed
A rubric for what is yours to say, what the model is allowed to draft, and what is allowed to ship unread. Voice tests. Boundary tests. A weekly receipts review against your own prior writing so drift is caught in the same week it starts. The work becomes faster without becoming anonymous.
- operator voice card
- ship-read boundary
- drift triggers
- weekly receipts review
what you walk out with
Twelve months from now, the people closest to you can still tell when you wrote something and when you didn't — and the things shipping under your name are recognisably yours. Your speed has gone up, but your authorship has not been dissolved into the statistical average.
02 systems literacy you can name what your stack is doing while it does it.
the failure
Eight agents under management. None of them surprise you in the demo. All of them surprise you on Friday afternoon when the customer asks why the proposal contradicts the SOW. The stack did not fail because it was dumb. It failed because no one owned the whole system.
what gets installed
A working map of inputs, outputs, failure modes, and drift for every agent in your stack. Standing post-mortems for every agent action that touched a customer. A monthly stack review against the map — what changed, what failed, what got promoted. The map becomes the thing everyone points at before blaming the model.
- agent stack map
- responsibility ledger
- failure-mode table
- promotion review
what you walk out with
When your stack does something unexpected, you can locate the cause in under thirty minutes — instead of three days of forensic Slack. Your team stops asking what did the AI do? and starts asking which boundary produced this?
03 coordination you manage agents the way you manage senior humans.
the failure
You delegate to the agent the way you would delegate to a chatbot. You get the work product of a chatbot. You then redo it. You then wonder why your week looks like the week before you had the agent. The agent becomes a second inbox instead of leverage.
what gets installed
Context briefs, escalation criteria, post-action reviews, standing rhythms. Treat each agent as a contractor with a JD, not a tool with a prompt. The rubric: would a senior human in that seat have shipped this? If no, the agent doesn't either. The stack gets managed like a team, not babysat like a toy.
- agent jd
- context brief
- escalation thresholds
- post-action review
what you walk out with
Six months from now, your throughput is 3-5x what it was, and the work shipping from the stack passes the same bar as the work shipping from your senior humans. You stop celebrating agent output and start governing agent performance.
04 navigation you rewrite the map in the same week the territory shifts.
the failure
Quarterly OKRs. The territory rewrote itself in week three. You spent weeks four through twelve executing against a map that no longer matches the ground. You hit the OKRs. You lost the quarter. The danger is not missing the plan; it is faithfully executing a dead one.
what gets installed
A weekly cadence for revising the strategic map — what changed in the market, what changed in the stack, what changed in your own conviction. A standing rule: if your current quarter's plan would not survive a fresh look this Friday, it is replaced this Friday. Direction becomes a living instrument instead of a quarterly artifact.
- weekly map review
- constraint log
- conviction ledger
- replace-by-friday rule
what you walk out with
You stop being the operator who is right on paper and wrong in reality. Your team starts trusting your direction because your direction tracks the ground they're standing on. The plan becomes credible because everyone can see it changing when the world changes.
05 translation you hold both ontologies in one head.
the failure
Your board still talks org charts and roles. Your stack runs on graphs and agents. You translate, badly, between the two. Decisions get made in the wrong ontology. The board misses what's happening. The stack does what the board asked for, which was the wrong thing. The cost is not confusion. It is entire teams executing true statements from incompatible worlds.
what gets installed
A working translation table between the legacy ontology (functions, departments, roles, deliverables) and the new ontology (graphs, agents, contexts, artefacts). Updated monthly. Used in every board update, every team meeting, every customer conversation. You learn to land graph-native decisions in rooms that still budget, hire, sell, and govern through legacy categories.
- translation table
- board-update grammar
- team-meeting bridge
- customer ontology notes
what you walk out with
You stop being the operator who lost the room because half the room was on the wrong ontology. You become the operator who can land a graph-native decision with a function-native board, every time. You can tell the board the truth without flattening the graph — and tell the stack the truth without confusing the board.
the protocol · how the pillars get installed
Five pillars. Three weeks. One install.
The pillars are the substance. The install is the protocol that gets them installed against your live business. The matrix is the short version: each week installs a different part of the operator, and each week ends in an artifact that proves the discipline moved from language into practice.
build the stack
your agents inherit standards, context, escalation, and review
pillarssystems literacy · coordination
artifactai operating stack
launch the edge
the work ships as evidence, not as a prettier strategy doc
pillarstranslation · navigation (outward)
artifactlaunched edge + signed decision cards
The artifacts are the proof the pillars were installed. If there is no artifact, the pillar was not installed.
The pillars are not a course. They are the operating system. If you do not install them, something else will install itself in their place — and it will not have your name on it.
free / paid
What we give away. What we charge for.
- free — the pillars, the rubrics, the essays, the primers, the manifesto — The five essays walk each pillar. The five rubrics let you self-mark today. The primer surfaces which pillar you're leakiest on. The manifesto is the case for the room. None of it is gated. None of it asks for an email. You can run the curriculum on yourself, alone, for free, starting today.
- paid — three weeks of live install — The founding install is the only thing we charge for. The install is where the pillars get installed against your live business — your P&L, your stack, your customers — with the founders, in twenty-one days. Your operator graph, your AI operating stack, your launched edge. Plus twelve months of the app holding the install as living memory. Founding price $5,000. Standard price $10,000 after may 28.
- why both — The curriculum works alone — if your discipline is heroic. Most operators' discipline is not. The install is the institutional scaffolding that turns a curriculum-you-mean-to-run into a curriculum-that-actually-got-installed. Same standard. Higher install rate.
next step
Three weeks. Fifteen seats. June 1.
The founding install. Five pillars get installed across three weeks against your live business. Fifteen founding seats at $5,000. Standard price $10,000 after may 28. The install begins june 1. Once the room is full, the next cohort is a year away.
normally $10,000 · 15 founding seats · closes may 28 · begins june 1